...

HANDBOOKOf

NEW MEDIA

Social Shapingand Consequences of ICTs

Edited byLEAH A. LIEVROUW

and SONIA LIVINGSTONE

SAGE PublicationsLondon l Thousand Oaks l New Delhi

13

New Media and Organizingat the Group Level

A N D R E A B . H O L L I N G S H E A Dand

NOSHIR S . C O N T R A C T O R

Tr.t[litionally, smal l groups have been def ined byrI,t,SIrchcrs as collectives ranging from a minimumIII I:‘XI. and in most cases three, to a maximum of 15UI \(I rneinbcrs (cf. McGrath, 1984). Members of!rllops have interdependent goals, are acquaintedVII/~ and interact with one another and have a senseof hslon;ing. Recent developments in digital com-mlinication technologies have brought about amdlcal change in our collective notion of what con-Mules a group. Members of groups no longer needIO hc formally constituted or to be co-present (intime‘ or place) to collaborate, share information orjoo:rliLe. Ins tead new technologies fac i l i ta te thecrulion. inaintenance and dissolution of groupsXIIOII~ individuals who use different devices (sucha\ ~)hones. mobiles, laptops, personal digital assis-tani\) lo interact over one or more of a variety ofchamls (audio, video, text and graphics) offeredhy ~~eral forums (such as In ternet newsgroups ,onhne chat sessions via Instant Messenger, and cor-pclr;lir in t ranets) . These developments have trig-ptl a shift in conceptualizations of groups fromthe traditional notion of ‘same time, same place’ to‘any [inrc. anywhere’ and, some would argue apocry-phally. ‘all the time, everywhere’. In addition tothr physica l and tempora l cons t ra in ts , develop-meats in new media have also el iminated con-sIr;llnts on the size of groups. In traditionalface-to-face groups, the size of the group is likely tobc rrlatiwly small and its membership is by defini-~IOII :I closed set. This is also true for some geo-pphically distributed work teams that collaborateucmf communication technologies such as videoand ~oniputer confercncing. However, that is not

the case in many Internet-based newsgroups,where there are l i te ra l ly hundreds of par t ic ipants(Alexander et al., in press). These participants maycoalesce as a group because of a common ‘prac-t ice’ , such as co l labora t ing on the development ofa software program, or because of a common‘in te res t ’ , such as the i r concerns about the use o f‘ s w e a t s h o p ’ labour prac t ices or the i r in te res t indownloading a particular genre of music. As aglobal community of consumers and producers weare grappling with the opportunities and challengesof these new fluid ‘group forms’ of organizing.

As researchers, we are challenged to redefine thetheoretical and methodological apparatus to studyhow new media shape, and are in turn shaped by,the ways in which we organize in groups . Beforethe development of the World Wide Web and theInternet, research on groups with technological sup-por t was dr iven by three bas ic goals : to examinehow adequately new media could permit groups toovercome t ime and space constra ints , to evaluatethe impact of technologies on the range and speedof members’ access to information, and to evaluatethe impact of technologies on the groups’ task per-formance (McGrath and Hollingshead, 1994).Much of the theory and research addressed whenand how the structure, interaction and performanceof technologica l ly enabled groups were s imi lar toand different frotn face-to-face groups. As such, thefocus of this research was on examining the ways inwhich new media served to subs t i tu te and enlargecommunication among group members (Contractorand Bishop, 2000). With the surge in digital com-municat ion technologies , researchers s tar ted to

7

c

c

222 NEW MEDIA AND ORGANIZING

reckon with the idea that most technologicallyenabled groups were inherently different from face-to-face groups, and that they were worthy of studyas entities in their own right rather than simply to bebench marked against equivalent face-to-facegroups. Many of the premises of exis t ing theorieswere be ing cha l lenged by technologica l deve lop-ments and risked becoming less relevant at best, andobsolete at worst. Researchers are currently rethink-ing their definitions of groups and are developingnew theories to explain and predict their behaviour,and are designing new methods to study them.

This chapter examines the role of new media atthe group level of analysis. Its emphasis is on howtechnology shapes and is shaped by the behaviourof groups, rather than on issues relating to thedesign of hardware and software systems forgroup collaboration. The organization of thischapter reflects the evolution in theory and researchon groups and new media. As we shall see, thetheory and research also reflects our evolving de5nitions of ‘new media’ - starting with early experi-ments in teleconferencing (audio and videoconferencing) in the 197Os, and continuing withproprietary computer-mediated communicationsys tems in the 19SOs, the r ise of the Internet andthe web as ‘open’ communica t ion ne tworks in the199Os, and the ub iqu i tous , pe rvas ive and mobi lecommunication environment that ushers us into thetwenty-first century. The chapter begins with a briefdescription of an early, but influential, classificationof technologies that support group interaction. Thesecond and th i rd sec t ions examine the theory andempirical findings of research that investigated howtechnologica l ly enabled group co l labora t ions a resimilar and different from face-to-face collabora-tions. As will become evident, most of this researchwas conducted, or at least premised, on conceptual-izations of groups prior to recent developments in*@h,t&m!fx@~ .Th~~f~~~.,cP~~~~~~~~~

reconceptualization of groups that takes intoaccount the new forms of organizing enabled bynew media . This reconceptual iza t ion a l lows for amore f luid , dynamic and act ivi ty-based def ini t ionof groups and technology, and is drawn from a net-work perspective. It presents a knowledge networkapproach to the study of groups and technology.

A CLASSIFICATION OF TECHNOLOWEST H A T SU P P O R T GR O U P S

Collaboration among group members entails cogni-tive as well as emotional and motivational aspectsof communication. Group members transmit, receiveand store information of various kinds, from eachother and from various other sources. Theseexchanges were viewed as distinct functions carriedout by group members . Hence , not surpr is ingly ,

scholars conceptua l ized the technologies tha tsupport these functions to also be distinct. With aneye towards retrospective synthesis of research inthis area, McGrath and Hollingshead (1993; 1991)presented a classification system for communicationsystems based on the functional role of technologiesto support group collaboration. The four categoriesof the classification system are based on whetherthe technology: (1) provides within-group commu-nication (i.e. group communication support systemsor GCSS); (2) supplements information available tothe group or i ts members by information drawnfrom databases (i.e. group infonnation support sys-terns or GISS); (3) suppor ts communicat ion wi ththose outside the group (i.e. group external supportsys tems or GXSS); and (4) s t ructures group taskperformance processes and task products (i.e. groupperformance support systems or GPSS). The classi-fication system was developed in the early 1990swhen the World Wide Web was in i t s infancy. I twas later updated to include communication tech-nologies available on the Internet (Hollingshend.2001). While the classification system was devel-oped at a time when distinct technologies supportedthese different functions, it continues to be a viableframework to organize and examine contemporarytechnologies that typically support more than one (itthese four functions.

GCSS: Technologies that Mediate orAugment Within-Group Communication

The signature feature of GCSS is its ability to per-mit group members to communicate using net\media. In some cases GCSS may mediate commll-nicat ion among members spat ial ly separated frowneach other while they are communicating. Exampleswould include video conferencing, or ‘texting’‘ls~~,~~Q~,~~~e~~~~~~~~~.~~~ \i p %WJ , .~n&.er cx+:~

GCSS may augment face-to-face communiutwn _by the use bf overhead projectors for communicatinggraphics or document-shar ing sof tware over nst.worked computers. Some GCSS support asynchro-nous communication for group members interactingin different time periods; others require that groupmembers interact synchronously. As these c‘~anl.pies illustrate, GCSS vary in the communicatl~~nchannels that are available to group mcmberjvisual, auditory, text and graphics.

Most research on GCSS has been based on IIKpremise that the fewer modalities afforded bj I&-nologically mediated communication would ‘liltt!rout some of the cues in face-to-face communicnlion(Culnan and Markus, 1987). Based on this as~mption, the research agenda sought to examine hog theperformance of groups using GCSS was moderatedby the particular task(s) and activities in which rhegroup was engaged, the experience of tht goupwith the technology, and the degree to which pro\~p

ORGANIZING AT THE GROUP LEVEL 223

Table 13.1 A typology ofgroup communication support systemsModalities available Svnchronous AsvnchronousVisualA u d i o

Text, graphics

I n t e r n e t

Video conferencePhone conference

Computer conference

N e w s g r o u p sChat rooms

Videocassette exchangeVoice mailAudiotape exchangeE-mailFax

Home pages, websites

members have a shared conceptualization of management programs that organize schedules ,relative expertise (Hollingshead, 1998a; 1998b; files, contacts and other information on desktops toHollingshead et al., 1993). In addition to examining facilitate information exchange with other members.the performance of groups using GCSS, some Microsoft Outlook, which comes preloaded onresearch has focused on the interaction process many PC-compatible computers, is one suchamong group members. This research (McGrath information management program. More recentand Hollingshead, 1994) has found evidence that examples include software agents such as ‘web-the sequencing, synchrony and timing of messages bots’, or web-based robots, that assist groupamong group members using GCSS is moderated members by providing them, in some cases pro-by the size and nature of the groups, as well as the act ively , wi th re levant informat ion scanned f romlevel of ambiguity among group members. digital repositories.

Table 13.1 provides examples of GCSS organi-zed by the communication channels provided by thetechnology (video, audio, text/graphics) and thetemporal distribution of members, i.e. whether theyare communicat ing synchronously or asynchro-nously. As noted at the start of this section, GCSScan support communication between members whoare co-present or are geographically distributed.However, as we shall see in the review of empiricalresearch, the preponderance of research on GCSShas been among geographically distributed groups.Culnan and Markus (1987) a rgue tha t th i s b iasref lec ts an ear ly preoccupat ion wi th the ro le ofGCSS to mediate rather than to augment face-to-face communication. The organizing scheme alsoinc ludes ca tegor ies for In te rne t technologies ,although World Wide Web browsers can supportvideo conferencing, audio conferencing and docu-m e n t s h a r i n g o n t h e I n t e r n e t .

GXSS: Supporting ExternalCommunication

GISS: Supplementing InformationAvailable to the Group

The GXSS funct ion is a special case of both theGCSS function and the GISS function. Com-munication between group members and key exter-na l ‘human’ agents can be done wi th any of theGCSS systems described above. At the same t ime,one can consider interaction with non-humanagents (such as webbots) external to the group asaccessing yet another kind of information database,thus making it a special case of GISS. Organi-zations are increasingly able to interconnect seam-lessly the human agents and non-human agents ontheir intranets with those of their clients, partners,suppl ie rs or subcont rac tors , v ia secure web-based‘extranets’ (Barr e t a l . , 1998) . As such, extranetsserve as a unified infrastructure for GXSS thatreaches beyond the traditional organizationalboundary or its digital analogue, the corporate‘firewall’.

(iroup members have access to many repositoriesof information or knowledge besides other groupmembers . These reposi tor ies include databases ,archives and intranets. Intranets are web-basedtechnologies that support knowledge distributionamong networks of teams within organizations. Thetypes of knowledge that are available to groupmembers on intranets can include: (1) humanresources , (2) sa les and market ing ac t iv i t ies ,I!) financial information, and (4) design and manu-facturing specifications and innovations (Bar et al.,1098). Other examples of GISS are information

GPSS: Modifying the Group’s TaskPerformance

For several decades, researchers have designed andeva lua ted s t ra teg ies to s t ruc ture the in te rac t ionamong group members in order to enhance theireffect iveness. These strategies, often under thestewardship of a facilitator or supervisor, constrainand structure the communication, the task informa-tion available, and/or the form and sequence of taskresponses permit ted and required of the group

2 2 4 NEW MEDIA AND ORGANIZING

Some examples of such strategies are brainstorming,the Delphi method and the nominal group technique(NGT) (for a summary, see McGrath, 1984) .

More recent ly , technologica l ly enabled groupperformance support systems (GPSS) have beendeployed to assist with these strategies. An influentialeffor t has focused specif ical ly on technological lyenabled strategies to enhance decision-makingamong groups. These GPSS are also cal led GDSSor group decision support systems (see Jessup andValacich , 1993, for d iscuss ion) . In the la te 1980sand early 1990s most GPSS were in the form ofdecision rooms - specially equipped computer labssuppor t ing synchronous g roups wi th co - loca tedmembers . Most groups used these systems to aug-ment their face-to-face decisions. These systemsvar ied as to the type of t ask suppor t p rov ided togroups, the size of groups that could use thesystem, and whether a t ra ined faci l i ta tor was nec-essary to augment the GPSS. Those that provideddirect task support for groups usually incorporatedan array of ‘modules’ , each of which s t ructures adifferent subset of a group’s tasks or different por-t ions of the group process on a g iven pro jec t . Forexample, a GPSS might include tools or modulesfor electronic brainstorming; for structuring variousforms of evaluation and voting (rating, ranking,weighing, pick one, pick any, etc.); for identifyingstakeholders and bringing their assumptions to thesurface; or for exchanging anonymous or identifiedcomments on any or al l topics. Efforts are underway to develop these systems to support asynchro-nous and synchronous groups on the Internet. Morerecent ly , GPSS have been designed to encompassmore than just decis ion-making. Current effor ts inthe area of workflow management, enterpriseresource p lanning and computer-suppor ted cooper-ative work (discussed by Star and Bowker andothers elsewhere in this volume) underscore effortsto enhance group performance beyond simplydecis ion-making.

THEORETKAL PERSPECTIVES

Most pr ior theory and research have focused pr i -marily on how groups using technology accom-plished their tasks differently from groups withoutaccess to technology. More specif ical ly , much ofthe early theory relevant to the study of groups andtechnology addressed how the interaction and per-formance of groups that were separated in spaceand time differed from face-to-face groups. Thisresearch centred on those technologies classified asGCSS. One se t o f theor ies dea l t wi th the top ic ofmedia choice or media select ion: how people makechoices about different media to use in their com-munication with others. Another set dealt with thetopic of media effects: how technologies can impact

group interaction processes and group outcomes. :\th i rd s t ream of theoriz ing explored the intcrrda-tions between technologies and group interaction b!attempting to integrate the arguments ofl’~c~iby media choice and media effects thcorl\r\.Speci f ica l ly , adapt ive s t ruc tura t ion theory (XT)examined how the s t ructures that are imp& 11)technologies recursively shape and in turn ;ircshaped by group interaction. Most of the empmalinves t iga t ions of th i s perspec t ive were conducr~tlwith technologies classified as GPSS. Finally. thi:most current theory that relates to groups .III~technology deals with the complexity of gr”Llpprocesses, and suggests that technology is onl) cinrof many factors that can influence group procc\\c,and outcomes.

Media Choice

Some of the earliest theoretical work on IWII;Ichoice was conducted before computer use WI\widespread , and hence deal t wi th cornrnunicall~lllsystems other than computers . Short e t a l . (1’17!11proposed the sociul presence model to prctlwwhich media individuals will use for certain ~IIWof interactions. Social presence refers to the depccof salience of the other person involved in the itIler-action, and was therefore assumed to be an ‘ohlcc-tive’ dimension that could be calibrated b! , Iresearcher independent of the users. They hyporhc-s ized that media differed in their social preseilcc.and that individuals are aware of and agree on 11~difference and use it as a basis of their mdiachoice. For instance, they argued that on an oqlcc-t ive scale , text-based communicat ion has a loacrsocial presence than video conferencing, which illtu rn has a lower soc ia l presence than face-to-l:~acommunicat ion. Further they argued that indivltlll-als would select a communication medium that JIXIa social presence commensurate with the task tl~!were t rying to accomplish. Specif ical ly , they TIC-dieted that individuals avoid a given medium fisr , \given type of interaction if they perceive hatmedium as not providing a high enough degree 111socia l presence for that type of in teract ion. TIK!also predicted that communication using media 1~in socia l presence would be more appropr ia te iilrtask-rela ted communicat ion while media high IIIsocial presence, such as fact- to-face communi~.i-t ion, were more appropriate for t ransact ing inri,r-personal (or socioemotional) content,

Daft and Lengel (1986) extended the idc.i\embodied in the social presence model in th,:rrtheory of media richness. They proposed that (III-ferent forms of communication differ in the ‘ritll-ness’ of the information that they provide, Richnc\\was defined as the abi l i ty of a medium to provltlcmultiple cues (verbal and non-verbal), and immed-ate (or quick) feedback, us ing mult iple modalin?\

ORGANIZING AT THE GROUP LEVEL ,>i--_

(text, video, audio and graphics). Based on thesecriteria they arrayed the various media from verylean (company policy manuals for rules and regula-t ions) to lean (formal information systems) tosomewhat rich (direct contact) to very rich (groupmeet ings) . Fur ther , they argued that the var iousinfonnation processing tasks conducted by groupmembers could also be objectively arrayed in termsof their equivocality and uncertainty. Some com-munication tasks, such as finding the latest salesfigures, entailed reducing uncertainty (that is, find-ing the right answer to a question). Other tasks,such as crafting a sales strategy, required reducingequivocality (that is, determining what is the rightquestion to answer). Media richness theory pro-posed that ‘rich’ media were more appropriate toreduce equivocal i ty and ‘ lean’ media were moreappropriate to reduce uncertainty. Daft and Lengelargued that managers use (and should use) differentcommunication methods of appropriate degrees ofrichness to deal with situations that differ in equivo-cality and uncertainty. Hence, different communi-cat ion media , or s t ructural mechanisms in theirterminology, need to be used for different types oforganizational tasks. The more equivocality a situ-ation involves, the richer the information requiredto deal wi th i t . They presented seven s t ruc tura lmechanisms ordered along an information richnesscontinuum based on capacity for resolving equivo-cality versus reducing uncertainty. The sevenmechanisms included: group meetings, integrators,direct contact , planning, special reports , formalinfonnation systems, and rules and regulations.

At the time media richness theory was first pro-posed, e-mail was not widely available in organiza-t ions ; however , th i s theory was fea tured qu i teprominently in early empirical research thataddressed predictors of e-mail usage in organiza-tions. It was argued that managers whose choice ofmedia reflected the equivocality or uncertainty ofthe task were perceived to be more competent .Some researchers (Trevino et al., 1990) found sup-port for this argument, but many others did not (e.g.El-Shinnawy and Markus, 1997). One of the earlycriticisms of the model was that, like social pres-ence theory, it assumed that media richness wasconsidered to be an objective dimension; that is,each medium provided the same amount of rich-ness, predetermined by the inherent attributes of thetechnology, regardless of who was using it (Culnanand Markus, 1997). Other scholars proposed thatmedia r ichness was a subject ive dimension. Forexample, e-mail may be perceived as a richermedium by people experienced with that techno-logy than by those who are not. Still others notedthat most tasks involved varying degrees of uncer-tainty and equivocality and that it was often not fea-sible to parse the task into subtasks that wereunifonnly high or low in terms of their uncertaintyor equivocality. As such, for these unbundled tasks

it did not make much sense to dictate the use of leanor rich media.

Social presence theory and media richness theorywere influential early attempts to understand mediachoice among group members. The lack of consis-tent empirical support for these theories was attri-buted to the theories’ assumptions about ascribingobjective attributes (social presence or media rich-ness) to different communicat ion technologies . Asa resul t , a l ternat ive media se lect ion theor ies wereput forward tha t could account for these incons is -tent findings.

One such theoretical formulation was the socialirEfluerlce model. Fulk et al. (1990) contended thatthe media r ichness model is more normative thandescriptive of communication patterns in organiza-t ions . They a rgued tha t ind iv idua l percep t ions ofthe information richness of various media can vary,and that it was important to measure those percep-tions rather than to rely solely on an objectiveassessment. They contended that objective featuresof media r ichness can and do inf luence individualpercept ions of media r ichness , but there are othersources of such influence, such as social inter-action. Drawing upon earlier research on sociallearning theory and socia l informat ion process ingtheory , they a rgued tha t soc ia l in te rac t ion in theworkplace shapes the creation of shared meanings.and that those shared meanings provide an impor-tant basis for shared pat terns of media se lect ion(Fulk et al., 1990; Schmitz and Fulk, 1991).

The social inf luence model hypothesized thatmedia perceptions and use: (1) are subject to socialinfluence; (2) may be subjectively or retrospec-t ive ly ra t ional ized; (3) are not necessar i ly a imedat maximizing efficiency: and (4) may be designedto preserve or create ambiguity to achieve strategicgoals. Schmitz and Fulk (1991) found that per-ceived (as distinct from objectively defined) e-mailrichness predicted individuals’ e-mail assessmentsand usage and that the opinions of colleaguesinfluenced others’ media assessments. These resultssuppor ted the not ion tha t o ther group memberscan inf luence how indiv idua ls perce ive and usetechnology.

The social influence model of media select ionexplici t ly recognized the role of group members’communicat ion networks in shaping thei r percep-t ion of media r ichness . An important impl icat ion,not addressed by the socia l inf luence theory , washow media se lec t ion in turn inf luenced the subse-quent structure of the communication network itself(Contractor and Eisenberg, 1990). For instance,group members may be socially influenced by othermembers in their primarily face-to-face communi-cation network to begin using e-mail. However.once these members begin to use e-mail , the newcontacts avai lable through this new medium mayenlarge and possibly modify their pre-existing com-munication network. That is, it is possible that the

2 2 6 NEW MEDIA AND ORGANIZING

networks that socially influence individuals’ mediachoices may in turn occasion a restructuring in theircommunicat ion network. In essence, th is observa-tion points to a ‘media effect’ resulting from a‘media choice’. The following section describes aninfluential stream of research on the effects ofmedia use on groups.

Media choice theories may be rendered less rele-vant today by developments in technologies.Increas ingly , the convergence to a uni f ied mul t i -modal (audio, video, text and graphic) forum forcommunicat ion makes in teres t in the d is t inc t ionsbetween media, and hence the quest ion of mediachoice, obsolete. Unlike the context in which mediaselection theories were developed, today it isincreasingly plausible - even probable ~ for groupmembers to simultaneously communicate viamultiple modalities through a single device. Anexample would be the use of the web page tosimul taneously communicate v ia audio and video,whi le shar ing a document , and jo in t ly execut ing agraphic s imula t ion .

Media Effects

Hiltz and Turoff (1978) were among the first todescribe differences between face-to-face andcomputer-mediated interaction in terms of social andpsychological processes, and to discuss the impor-tance of task-media contingencies. Hiltz and Turoffargued that groups communicat ing via computerhad access to a narrower band of communicat ionthan groups communicating face-to-face. For exam-ple , non-verbal communicat ion and paralanguageeither were not available or were substantiallyreduced in computer-mediated communication. Insome s i tuat ions , such narrowband communicat ionallowed information to be communicated with moreprecision and less noise, and afforded the opportu-nity for rational judgement processes to operate inthe group with less intrusion of non-rational consid-erations. In other situations, computer conferencingneeded to be supplemented by other media in whichnon-verbal communicat ion and paralanguage wereavailable. They were also among the first to presentempirical findings that explored the effects of com-puter conferencing on the distribution of participa-t ion among members, on the amount of task andsocial communication, and on user responses to theava i lab i l i ty and the i r sa t i s fac t ion wi th the sys tem(Hiltz et al., 1986).

Kiesler et al. (1984) provided a theoretical ration-ale as to why and how groups will differ when theyuse computer-mediated as compared with face-to-face communication. They proposed that computer-mediated communication depersonalizes thein te rac t ion process , with several concomitanteffects. Individuals tend to lose mental sight of theirin teract ion par tners . At the same t ime, they lose

access to a variety of cues that provide feedback tomembers regarding the impact of their behaviour oninteraction partners, their status and their indivi-duali ty. Thus, computer-mediated communicationremoves substantial social infomlation and elimi-nates much of the feedback that people ordinarilycommunicate to one another face-to-face. This canhave both positive and negative influences on theinteraction processes, task outcomes and responsesof users (Sproull and Kiesler, 199 1).

People feel less inhibited when interactingthrough a computer network as a result of the reduc-tion in social cues that provide information regard-ing one’s status in the group. Therefore, participantsconcentrate more on the messages and less on thepersons involved in the communication. Individualsfeel less committed to what they say, less concernedabout it, and less worried about how it will bereceived by their communication partners. Becausepeople communicating electronically are less awareof social differences, they feel a greater sense ofanonymity and detect less individuality in others.As a consequence, individuals engaged in computer-mediated group interaction tend to:

1 feel more anonymous and detect less individual-ity in their communication partners;

2 par t ic ipate more equal ly (because low-s ta tusmembers are less inhibited);

3 focus more on task and instrumental aspects andless on personal and social aspects of interaction(because the context is depersonalized);

4 communicate more negative and more uninhibi-ted messages (because they are less concernedwith politeness norms that tend to regulate com-munication in face-to-face groups); and

5 experience more difficulty in attaining groupconsensus (both because of elimination of muchinterpersonal feedback, and because of reducedconcern with social norms).

All of these effects have been demonstratedempir ical ly ( for review, see Kiesler and Sproul l ,1992), and will be revisited in greater detail later inthis chapter.

McGrath and Hollingshead (1993; 1994) build-ing on the work described above and applying it towork groups, maintained that group interaction andperformance are greatly affected by the type anddifficulty of the task that the group is performing,and that the effects of technology on group inter-act ion and performance interact wi th task type.They hypothesized that the effectiveness or5 group -on a task will vary with the fit between the richnessof the information that can be transmitted using thatsystem’s technology and the information richnessrequirements of the group’s task. However , asgroups developed more exper ience wi th a givencommunicat ion technology, the r ichness of theinformation that could be transmitted effectively VI,Ithat technology would increase.

ORGANIZING AT THE GROUP LEVEL 337

McGra th and Hol l ingshead pos i t ed tha t g rouptasks differed in their information richness require-ments. Information richness referred to how muchthe information contains surplus emotional, attitu-dinal, normative and other meanings, beyond theliteral cognitive denotations of the symbols used toexpress it. They also posited that communicationmedia differed in the r ichness of the informationthat they can and do convey. Face-to-face commu-nication among interpersonally involved humanswas the richest medium; communication in writtenform among strangers was the least rich. Computercommunication among group members inexperi-enced with the technology is at the low-richnessend of that continuum.

Drawing from McGrath’s (1984) task typology,McGrath and Hol l ingshead hypothes ized tha tgroups working on genera te tasks (e .g . s implebrainstorming tasks) do not require the transmissionof evaluative and emotional content. As a result,computer-supported groups may brainstorm moreeffectively than face-to-face groups. At the otherend of the cont inuum, groups negot ia t ing andresolving conflicts of views or interests may requirethe transmission of maximally r ich information,including not only ‘facts’ but also values, attitudes,emotions, etc. As a result, groups interacting face-to-face should perform such tasks more effectivelythan groups interacting via computer. In betweenthe two ends of the continuum are intellective tasksthat have a correct answer or decision-making tasksthat do not have a correct answer , which mayrequire some intermediary level of informationr ichness . The predic t ions for generate tasks andnegotiation tasks received empirical support(Gallupe et al., 1991; Hollingshead et al., 1993;Valacich et al., 1994), but not those for intellectiveand decision-making tasks (Hollingshead et al.,1993; Straus and McGrath, 1994).

McGrath and Hollingshead (1994) also predictedtha t communica t ion technologies could provideinformation of increasing r ichness over t ime, asgroups learned how to embed additional emotional,attitudinal, normative and other meaning throughc o n t i n u e d e x p e r i e n c e .

In summary, the theoretical arguments reviewedin this section offer three related perspectives onhow technologies may influence the processes andoutcomes of groups. While they vary in their levelsof sophistication and theoretical complexity, allthree theoretical approaches to media effects arebased on the premise that technological attributesof different media influence key aspects of theinteraction process. These key aspects include theavailability of non-verbal cues, the potential foranonymous contributions, the ability to communi-cate status differentials, and the information richnessof the medium. These key aspects in turn helped orhindered the group’s interaction process (such asamount of participation, distribution of participation

and negat ivi ty in communicat ion on ‘ f laming’) , aswel l as the g roup’s ou tcomes ( such as consensus ,accuracy and speed of decision-making).

As such these theoretical perspectives on mediaeffects acknowledge a modicum of technologicaldeterminism. Not unl ike the media choice theoriesof socia l presence and media r ichness , d iscussedin the previous section, the theories of mediaeffects described in this section do not privilege asocially constructed explanation for understandingmedia effects. The following section offers a theo-retical framework that explicitly recognizes thesocial nature of technology and advocates an inex-tricable interrelatedness between media choice andmedia effects.

Adaptive Structuration Theory

Adapt ive strucmration theory CAST), p roposed byPoole and DeSanctis (1990) and inspired by theinf luent ia l theore t ica l cont r ibu t ions of Giddens’(1984) s t ruc tura t ion theory , s t resses the impor-tance of group interaction processes, both in deter-mining group outcomes and in mediating theeffects of any given technology. Essentially, asocial technology presents a structure of rules andoperations to a group, but the group does not pas-sively choose the technology in its pre-existingform. Rather , the group act ively adapts the tech-nology to its own ends, resulting in a restructuringof the technology as it is meshed with the group’sown interaction system. Thus, a technology can bethought of as a set of social pract ices that emergeand evolve over t ime.

From this point of view, the structure of a groupis not a permanent, concrete set of relationsbetween members and their tasks. Rather, the struc-ture is an evolving set of rules and resources avail-able to them to produce and reproduce the apparentlys table in terac t ion sys tems tha t we observe . Thus ,there is a recursive process between the structures(or the rules and resources in a group) and thesystems (the interaction patterns in the groups). Therules or resources in the group shape the in terac-t ions pat terns among group members. The interac-t ion pat terns among the group members , in turn ,reify or subvert the rules and resources in thegroup. This recurs ive process i s ca l led adapt ivestructuration.

The ru les and resources tha t groups use in thestructuration process are sometimes created on thefly by the group, but more often they arejU$ul/~~uppropriated by the group based on the social con-text in which it is embedded. Appropriation is theprocess by which a group selects features of a tech-nology and socially constructs their meaning. It isthrough such appropriation that a group can chooseto use a new technology. In some cases the groupmay not appropriate a technology in ways that were

2 2 8 NEW MEDIA AND ORG,3NIZING

in tended by the des igners of the technology . Thiss i tua t ion i s refer red to as an irattic uppt.opriatiotz.For instance, a group may have access to a groupdecision support system (GDSS) that provides themwith an opportunity to vote on their ideas. The votingtool is intended by the designers of the technologyto faci l i ta te democrat ic del iberat ion among groupmembers. However, in some instances members ofa group may use the voting tool to prematurelyclose off discussion of an issue. This action wouldi l lus t ra te an i ronic appropr ia t ion of the GDSS. Byfaithfully or ironically appropriating a technology,each group invests meaning in, and thereby adaptsfor i t s use , the ru les and resources tha t i t d rawsupon. Both technology and context af fect groupprocesses and outcomes because they affect thisappropriation process.

Empir ical research has shown that different , butseemingly similar, groups appropriate the sametechnology in different ways (DeSanctis and Poole,1997; Poole and DeSanctis, 1992; for a review seeDeSanct i s and Poole , 1994) . Zack and McKenney(1995) offer a recent example of work in this tradi-t ion They examined the appropriat ion of the samegroup authoring and messaging computer system bythe managing editorial groups of two morningnewspapers owned by the same parent corporation.Drawing upon Poole and DeSanctis’ (1990) theoryof adapt ive s t ruc tura t ion , they d iscovered tha t thetwo groups’ appropr ia t ion of the technology , asindexed by their communicat ion networks, differedin accordance with the different contexts at the twolocat ions , Fur ther , they found evidence tha t thegroups’ performance outcomes for s imilar taskswere mediated by these interaction patterns.

Adaptive structuration theory continues to be anincreas ing ly in f luen t ia l perspec t ive to unders tandthe socially constructed ways in which groups’choice of media and the effects of media on groupscoevolve. It provides a powerful analytic frame-work to account for stability and change in agroup’s appropriation of new media. While the util-i ty o f a s t ruc tura t iona l perspec t ive to the s tudy ofgroups’ use of new media is compelling, there con-tinues to be a debate about the extent to whichempir ica l s tudies of fer a ‘ tes t ’ as opposed to anillustration of structuration theory’s ability toexplain the unfolding of complex processes(DeSanctis and Poole, 1994). Indeed, in a review ofempirical studies from a structurational perspective,one would be hard pressed to identify a single workwhich failed to find support for adaptive structura-t ion theory. Such overwhelming endorsement ofa theory be l ies an under ly ing concern about thepotential falsitiability of the theory. An appropriatechallenge therefore would be to come up with spe-cif ic predict ions from the theory that , i f they werenot empirically validated, would plausibly representa refutat ion of the premises of adaptive s tructura-t ion theory . Complexi ty theory , d iscussed in the

next section, offers a novel and useful approach totranslate the richly evocative, but highly abbrevi-ated, verbal explications of adaptive structurationtheory into precise, falsiliable hypotheses that canbe empirically validated (Poole, 1997).

Groups as Complex Systems

In the pas t decade there has been a p le thora ofscholarship calling for the extension of complexitytheory - arguably a mainstay of many disciplines inthe physical and life sciences ~ to social sciences ingenera l , and to the s tudy of groups in par t icu lar(Arrow et al., 2000; Contractor and Seibold, 1993;Contractor and Whitbred, 1997; Gersick, 1991:McGrath, 1991). The motivation for this call stemsfrom a widely shared frustration with extanttheor ies , which have proven to be inadequate a tuntangling with precision the complexity in groupprocesses . The phenomena descr ibed in verbalexpos i t ions of , say , adapt ive s t ruc tura t ion theoryinvoke a multitude of factors that are highly inter-connected, often via complex, non-l inear , dynamicrelationships. Lamenting the failed promise ofear l ie r forays in to sys tems theory , Poole notes ,‘Most often, systems theory became a metaphor,rather than an instrument of analysis’ (1997: 50).Two streams of research that attempt to go beyondthe use of complexity theory as a metaphor(Contractor , 1999) have been developed to dealwith the complexi ty of groups’ use of new media:groups as self-organizing systems (Contractor andSeibold, 1993; Contractor and Whitbred, 1997) andgroups as complex, adaptive and dynamic systems(Arrow et al., 2000).

Groups as Self-organizing Systems

In general terms, ‘sel f -organizing systems theory(SOST) seeks to explain the emergence of patternedbehaviour in systems that are initially in a state of dis-organization. I t offers a conceptual framework toexplicitly articulate the underlying generative mech-anisms and to systematically examine the processesby which these mechanisms generate , sus ta in andchange existing structures or elaborate new struc-tures’ (Contractor and Seibold, 1993: 536).

llya Prigogine and his colleagues proposed thetheory of self-organizat ion. In an effort that con-tributed to a Nobel Prize, Prigogine and hiscolleagues (Glansdorff and Prigogine, 1971)mathemat ical ly proved that systems that exhibi temergence of spontaneous order must meet thefollowing logical requirements:

1 At leas t one of the components in the sys temmust exhibit autocatalysis, i.e. self-referencing.

2 At leas t two of the components in the sys temmust be mutual ly causal .

ORGANIZING AT THE GROffP LEVEL 2 2 9

; The system must be open to the environmentwith respect to the exchange of energy andmatter .

4 The system must operate in a far-from-equilibrium condition.

These four requirements offer, at a very abstractlevel, the conditions under which any system canself-organize. Our interests here are in applyingthese concepts to the study of groups using newmedia. Contractor and Seibold (1993) developed aself-organizing systems model for groups’ use ofgroup decis ion suppor t sys tems (GDSS). Theydeveloped a model based on the theoretical mechan-isms speci f ied by adapt ive s t ruc tura t ion theory[Poole and DeSanctis, 1990; discussed in the previ-ous section) about the recursive interrelationshipbehveen the s t ructures ( the ru les and resourceswithin the group) and the systems (the interactionpatterns among the group members) . Contractorand Seibold (1993: 537-8) specified four genera-tive mechanisms that were consistent with the theo-retical tenets of adaptive structuration theory andmet the logical requirements of self-organizingsystems theory:

1 Members’ expertise (or resources) with thetask wi l l re inforce the content and pa t te rnof their communicat ion during GDSS-baseddiscussions.

2 The content and pattern of members’ communi-cation will reinforce their perceptions of thegroup’s norms for structuring the GDSS-baseddiscussion.

3 Members expertise (or resources) withGDSS wil l re inforce their percept ions of thegroup’s norms for structuring the GDSS-baseddiscussions.

4 Members’ perceptions of the group’s normsfor s t ruc tur ing the GDSS-based d iscuss ionwill reinforce the content and pattern of theirc o m m u n i c a t i o n .

Using simulations, they showed that based on thesefour theoret ical mechanisms the group’s use ofGDSS would self-organize only under a very speci-ficrange of initial conditions. A group using GDSSwas considered to have self-organized when thegroup’s structures (that is, its members’ perceptionsof the rules) were stable and the group members’interaction patterns were reproducing and reinforc-ing (rather than subverting) these stable structures.The simulation also provided precise conditionsunder which the groups would 1~01 successfullyappropriate the technology. That is, the group mightinitially attempt to use the technology but wouldthen discontinue its use. These results, theoreticallygrotmded in adaptive structuration theory and logi-ca l ly consis tent wi th se l f -organiz ing sys temstheory, represent plausible occurrences in groups’use of new media. They also respond to one of the

criticisms levelled against adaptive structurationtheory by making i ts explanat ions more amenableto falsification. In general terms, the approach illus-trates how self-organizing systems theory can offerthe logica l condi t ions and the analy t ic f rameworkto discover precise, empirically falsifiable hypothe-ses about the use (and lack thereof) of new mediaby groups.

Groups as Complex, Adaptiveand Dynamic Systems

Arrow et al. (2000) have proposed a general theoryof complex systems, which embeds technology asone aspect of the system. This theory builds on thetime interaction and performance (TIP) theory pro-posed by McGrath ( 199 1). TIP theory assumes thatgroups pursue multiple functions for multiple pro-jects by means of complex time/activity paths.Arrow et al. (2000) extend this theory by proposingthat all groups act in the service of two genericfunctions: (1) to complete gr’oup projects and (2) to,filfil me&el needs. A group’s success in pursuingthese two fimctions affects and depends on the via-bility and integrity of the group as a system. Thus,mnifztclinillg system ilztegrig, becomes a third func-tion, instrumental to the other two. A group’ssystem integr i ty in turn affects i t s abi l i ty to com-plete group projects and fulfil member needs.

Groups include three types of elements:(1) people who become group rnel,lDers; (2) goals thatare embodied in group projects: and (3) resourcesthat get transformed into group technologies.Technologies differ in how much they faci l i ta te orcons t ra in in te rpersona l ac t iv i ty , t a sk ac t iv i ty andprocedura l ac t iv i ty ; and in how effec t ive ly theysupport di f ferent ins t rumental funct ions ( i .e . pro-cessing of information, managing of confl ict andconsensus, and motivation, regulation and coordi-nation of member behaviours).

A group pursues i t s func t ions by c rea t ing andenacting a coordinated pattern of member-task-tool relations, its coor-dinntion netw0r.k. Theful l coordina t ion ne twork inc ludes s ix componentne tworks : (1 ) the member network, or pa t te rn ofmember-member relations (such as status rela-tions); (2) the tnsk network, or pattern of task-taskrelations (e.g. the required sequence for completionof a set of tasks); (3) the tool netwol.k, or pattern oftoo l - too l re la t ions (e .g . the procedure by which atechnology can be used most eff ic ient ly) ; (4) thelabour network, or pattern of member-task rela-tions (i.e. who is supposed to do what); (5) the rolenetwork, or pattern of member-tool relat ions ( i .e .how members do their tasks) ; and (6) the job net-walk, or pattern of task-tool relations (e.g. whatpiece of equipment must be used for a given task).

The l ife course of a group can be characterizedby three logically ordered modes that are conceptu-a l ly d is t inc t but have fuzzy tempora l boundar ies : 3

2 3 0 NEW MEDIA4 AND ORGANIZING

jbrmation, operation and metamorphosis. As agroup forms, people, intentions and resourcesbecome organized in to an in i t ia l coord ina t ion net-work of relations among members, projects andtechnology that demarcates that group as abounded soc ia l en t i ty . As a g roup opera tes overt ime in the service of group projects and memberneeds, i ts members elaborate, enact , monitor andmodify the coordination network establishedduring formation. Groups both learn from theirown exper ience and adapt to events occurr ing intheir environment . I f and when a group undergoesmetamorphosis , i t d issolves or i s t ransformed intoa different social entity.

O V E R V I E W OF M A J O R E M P I R I C A L F I N D I N G S

A number of scholars have written literaturereviews that examine communicat ion technologiesand groups (e.g. Benbasat and Lim, 1993;Hollingshead and McGrath, 1995; Kiesler andSproull, 1992: Kraemer and Pinsonneault, 1990;McGrath and Hol l ingshead , 1994; McLeod, 1992;1996; Seibold et al., 1994; Williams, 1977). Most ofthese reviews have compared the interactionprocesses and outcomes of computer-mediatedgroups with those of face-to-face groups. Several ofthose reviews have reached the same conclus ionsabout the s ta te of knowledge in th is a rea : namely ,that more theory-guided and programmatic researchis needed (e .g . Hol l ingshead and McGrath . 1995;McLeod, 1992).

Interaction Patterns

Many studies have revealed that groups interactingvia computers have more equal participation amongmembers than groups interacting face-to-face (e.g.Clapper et al., 1991: Daly, 1993; Dubrovsky et al.,1991; George et al., 1990; Hiltz et al., 1986;McLeod, 1992; Rice, 1984; Siegel et al., 1986;Straus, 1996; Straus and McGrath, 1994; Zigurset al., 1988). As described earlier, the general expla-nation for the effect is that people feel less inhibitedwhen in te rac t ing through a computer ne twork as aresu l t o f the reduc t ion in soc ia l cues tha t p rovideinformat ion regard ing one’s s ta tus in the group .Because people communicat ing electronical ly areless aware of social differences, they feel a greatersense of anonymity and detect less individuality inothers (Sproull and Kiesler, 1991). It is important tonote some common elements across this set of stud-ies. These studies were conducted during oneexperimental session with nd hoc groups consistingof s tudents in a l abora tory se t t ing . However , i t i salso important to note that this finding was observedacross a variety of communication technologies.

Many studies have also showed no evidence ofthe participation equalization effect in computer-mediated groups (Berdahl and Craig, 1996;Hollingshead, 1996b; Lea and Spears, 1991;McLeod and Liker , 1992; McLeod e t a l . , 1997;Saunders et al., 1994; Spears and Lea, 1992;Watson et al., 1988; Weisband, 1992; Weisbandet al., 1995). In fact, most showed that status differ-ences among participants were displayed in theirinteraction in the computer-mediated setting. Oneexplanation for the inconsistency of findings acrossstudies is that s tatus differences among memberswithin the groups may have been di f ferent ia l lysalient across studies. When members’ identitieswere known or were available visually, the statusdifferences in the number of contributions and theperceived inf luence of those contr ibut ions weremaintained in the computer-mediated setting. Whenthey were not or when members’ contributions wereanonymous, the par t ic ipat ion equal izat ion effectwas more likely to occur.

I t i s a l so poss ib l e tha t t he pa r t i c ipa t ion equal-iza t ion may be an indicat ion of how the mediumreduces the baseline of each member’s participa-tion rather than how the medium leads to increasedpar t ic ipat ion of low-sta tus members dur ing thegroup d i scuss ion (McGra th and Hol l ingshead ,1994; Spears and Lea, 1994). It takes more time totype a message on a computer network than it doesto say that same message verbal ly . In the experi-ments cited above, the computer sessions were atleast as long as those face-to-face group meetings;however, the amount and the rate of communica-t ion in the computer-mediated set t ing were muchless . Another poss ib le t echnolog ica l exp lana t ionfor greater egalitarian participation patterns incomputer-mediated settings is that electronicgroup members have the ability to participatewi thou t i n t e r rup t ion , s i nce t u rn - t ak ing i s no t anorm in a computer-mediated environment(Weisband e t a l . , 1995) .

A number of studies have found that computer-mediated groups exchange less information and areless likely to repeat information in their decisionsthan face- to-face groups (Holl ingshead, 1996a;1996b; McLeod et al., 1997; Straus and McGrath,1994) . In some cases , th is reduct ion can lead topoorer outcomes for newly formed groups (cf.Hol l ingshead , 1996a; 1996b) .

Performance

Very few s tudies have demonstra ted that groupscommunicating via computer perform better thangroups interacting face-to-face, although manyhave demonstrated that computer-mediated groupsperform less well than or equally well as face-to-face groups (for reviews see McGrath andHol l ingshead , 1994; McLeod, 1992; 1996) . Even

aCrtCti

It1L

;

ORGANIZING AT THE GROUP LEVEL 231

PSa nIYPs:e-n dten

though computer-mediated groups generate lesscommunication and use less information in theird&ions, they take longer to make themthollingshead, 1996a). They are also less likely toreach consensus (for reviews see Hollingshead andMcGrath, 1995; Kiesler and Sproul, 1992).

As described earlier, there seems to be an inter-action effect of task and technology on the qualityof group performance. Computer groups producemore ideas of higher quality on idea generationtasks . Face- to-face groups tend to have higher-quality products on intellective and negotiationtasks . However , i t may be the s t ruc ture tha t i simposed by the technology rather than the techno-logy itself that is responsible for this effect(Hollingshead and McGrath, 1995). The task struc-ture may include: procedures that s impl i fy thehandling of complex information; procedures thatexplicate agenda, thus making group process moreorganized; and procedures that expose conflict andhelp the group to deal wi th i t . Some researchshowed that a paper and pencil version of the taskstructure imposed by the technology (i.e. withoute lect ronic communicat ion) gave higher-qual i tydecisions than the same task structure provided by aGPSS, which in turn was higher than the no-structure face-to-face condition (Hollingshead andMcGrath, 1995; Watson et al., 1988). In some cases,newly formed groups on computers may have prob-lems with task structure that requires more complexinformation processing (Hollingshead, 1996a).

Longitudinal research comparing the impact ofcomputer-mediated and face-to-face communica-tion over time has brought into question previousfindings of significant differences in performancebetween face-to-face and computer-mediatedgroups. That research has shown that computer-mediated communication hinders the interactionprocess and performance of groups initially, butover time, groups can adjust successfully to theirmode of communication (see McGrath et al., 1993and Arrow et al., 1996 for overviews). In addition,work on the interpersonal and relationshipaspects of computer-mediated communication overtime complements this finding. Walther andBurgoon (1992) showed that members ofcomputer-mediated groups felt less connected toone another initially, but over time, members ofcomputer-mediated groups expressed more posi-tive feelings about one another that approximatedthose expressed by members of face-to-facegroups. The transient effects of technology werealso illustrated in a longitudinal study comparingthe developments of norms in groups using GDSSwith groups not using GDSS. Contractor et al.(1996) found that while members of non-GDSSgroups were initially more likely than members ofGDSS groups to socially influence one another’sperceptions of the group’s norms, this differencedissipated over time. That is, in the long term,

groups using GDSS were no more likely thangroups no t us ing GDSS to soc ia l ly in f luence oneanother’s perceptions of the groups’ norms.

T HE R E C O N C E P T U A L I Z A T I O N OF G R O U P S

AND NEW MEDIA AS KNOWLEDGE NETWORKS

While it should be evident that the study of groupsand new media is a vibrant area for research, wenow return to the opening statements of this chapterabout the theore t ica l and ana ly t ic cha l lenges tha tconfront scholars who consider the ways in whichthe ‘new’ new media of the twenty-f i rs t centurywill influence our ability to organize in groups. Inconclusion, we offer a reconceptualization ofgroups’ use of new media from a knowledgenetworks perspective.

From Knowledge Managementto Knowledge Networks

Knowledge management is a critical concern forcontemporary organizat ions , and i t i s expected tobecome increasingly important in the future(Nonaka and Takeuchi, 1995). It has long been rec-ognized that computers could increase the range,depth and speed wi th which informat ion could beacquired, processed, presented for use and sharedfor collaborative efforts. However, research in thisarea has given little attention to theoretical or con-ceptua l i ssues about informat ion acquis i t ion , pro-cessing and integration, and even less attention totheoretical issues about the antecedents and conse-quences of different pat terns of information dis tr i -bution within work groups, and the conditionsunder which information can be and is easily sharedamong group members. Recent developments intechnologies have shown their potential as knowl-edge management systems, although little is knownabout the social challenges and motivations forgroup members to use these systems effect ively.These challenges call for a knowledge networkapproach (Monge and Contractor, 2001) andknowledge-based theor ies to unders tand groups’use of new media

Groups as Knowledge Networks

The prol i fera t ion of d ig i ta l technologies has dra-mat ica l ly changed the nature of work in groups .These technologies , as descr ibed previous ly , havethe potential to provide many benefits to groups byl inking people who have common goals and in ter -ests but are separated in time and space. They mayenable organizat ions to develop effect ive teamsfrom workers who are geographical ly d is t r ibuted .

232 NEW MEDIA AND ORGANIZING

Today, in stark contrast to just a decade ago, virtualteams consider having employees located in t imezones far removed from one another (such asCalifornia, Ireland and India) as a competitiveadvantage rather than a disadvantage. Members ofdistributed work teams can work round the clock inorder to meet the competi t ive demands of a globalmarketplace. In some cases the members of theseteams are ‘e- lancers’ (electronic freelancers) whocoalesce on a short-term project and then disperse.In other cases, the technologies have the potential toenable the organization to hire and retain the bestpeople , regardless of locat ion (Townsend et a l . ,1996). These changes have led scholars to call for areconceptual izat ion of groups as much more f luid,dynamic, mult iplex and act ivi ty based (Goodmanand Wilson, 2000).

Clear ly these new technologies have the poten-t ia l to nur ture a team by l inking the members notonly to one another but a l so to a la rge number ofinternal and external knowledge repositories.Conceptually, therefore, it is increasingly useful toconsider the group and its members as a network ofagents, where some of the agents are human agentswhile others are non-human agents (such as knowl-edge reposi tor ies . avatars and webbots) . Humanagents communicate with one another by retrievingand allocating information relevant to their collec-tive tasks. An increasingly vexing question thatgroup members face in this networked environ-ment is not which medium to use (as was addressedby earlier theories of media choice), but ratherwhich agent to use.

Groups and the media they use can be useful lyreconceptualized as a lo~o~&~ge network. A net-work is made up of a se t of nodes and re la t ionsbetween these nodes . The nodes tha t conta in theknowledge can be people. databases, computerfi les or other forms of reposi tor ies . The relat ionsare the communicat ion re la t ions ( tha t i s , publ ish-ing, re t r ieving, a l locat ing) among the nodes . Thelocation of knowledge within this network ofagents can vary along a cont inuum from central-ized. where knowledge resides with only one agent,to distributed, where knowledge exists amongmany agents (Farace et al.. 1977). Distributedknowledge may refer to the parts of a larger knowl-edge base, each possessed by separate actors withinthe network. In this form of distributed knowledge,actors bring relatively unique. non-redundantknowledge which enables a collective to accom-plish complex tasks. Distributed knowledge occursat many levels in the empirical world, includingwork groups, large-scale project teams, andin te rorganiza t iona l s t ra teg ic a l l i ances . Al te rna t i -vely , d is t r ibuted knowledge may refer to the f lowor diffusion of knowledge, which increases thelevel of knowledge among all actors.

Communicat ion networks , ac tual knowledge ner-works, and cognitive knowledge networks are

different ways of conceptualizing the network ofagents . Communicat ion networks represent thedegree to which individual agents interact witho ther agents in the ne twork . Actua l knowledgenetworks represent the actual distribution ofknowledge among the network of agents. Cognitiveknowledge networks represent individuals’ percep-tions of the distribution of knowledge in the net-work of agents. Knowledge networks are dynamic,in terms of both agents and linkages. Agents join orleave a knowledge network on the basis of tasks tobe accompl ished, and thei r levels of in teres ts ,resources and commitments. The links within theknowledge network are also likely to change on thebasis of evolving tasks, the distribution of knowl-edge within the network, or changes in the agents’cognitive knowledge networks. New media, such asintranets, serve both as the infrastructure that supports the development of relations in the networkand as the nodes in the network. In our ownresearch, we have appl ied a knowledge networkperspective to theories that investigate new lnediause in g roups and o rgan iza t ions (Hol l ingsheadetal., 2001; Monge and Contractor, 2001).

We believe there is tremendous potential for thedevelopment and extension of theories which seekto expla in the development of a group’s use ofmedia as a knowledge network of human and non-human agents. The knowledge network perspectiveis especially well suited to test multiple theoriesand their contradictory or complementary inf lu-ences on the evolution of the groups. Knowledgenetworks and their defining characteristics can berepresented and analysed exceptionally well usingtechniques developed within the field of social net-work analysis (Wasserman and Faust, 1994). Thesetechniques enable researchers to examine thedynamics of relations at multiple sites and acrossdi f ferent levels of analys is ( indiv idual , dyads .group, organizations and industries). It is difficultto predict the types of configurations that groupswith technology will take in the future. Regardlessof their forms, a knowledge network perspect ivewill allow future researchers to examine, describeand evaluate new media and organizing at thegroup level.

REFERENCES

Alexander, S.C.. Wille, J. and Hollingshead, A.B. (inpress) ‘Support at your keyboard: a shady of on-line sup-port groups’, in L. Frey (ed.), Grwp Cbnrnwricnk~rr inCoute.r/, vol. 2. Hillsdale, NJ: Erlbaum.

Arrow, H., Berdahl, J.L.. Bouas, KS., Craig, KM.,Cummings, A.. Lebie. L., McGrath, J.E., O’Connor, K.M.,Rhoades, J.A. and Schlosser, A. (I 996) ‘Time. techno-logy. and groups: an integration’. Contpttter .Su,q~w/~~iCooperati~v W&k, 3: 253-61.

ORGANIZING AT THE GROUP LEVEL 233

irr~lw, Il., McCrath, J.E. and Berdahl, J.L. (2000)Small Groups as Complex Sv.Ftenrs: Formation,

(;wdinafion, Developmet~/. and Adaptation. ThousandOaks, CA: Sage.

Bar, F., Bonus, M. and Hanson, W. (1998) ‘Webportfolios: leveraging intranets, extranets, and the inter-net for strategic gain’. Stanford University working paper.

Bcnbasat, 1. and Lim, L. (1993) ‘The effects of group,task, context, and technology variables on the useful-ness of group support systems: a meta-analysis ofenperimental s t u d i e s ’ , Smcrll Group Resea~h,

14: 430-62.Berdahl, J.L. and Craig, K.L. ( 1996) ‘Equality of partici.

pation and influence in groups: the effects of communi-cation medium and sex composition’, ComputerSupported Cooper&w Work, 4: 179-202.

Clapper, D.L.. McLean, E.R. and Watson. R.T. (1991)‘An experimental investigation of the effect of groupdecision support on normative influence in smallgroups’, in J.I. de Gross, 1. Benbasat, G. DeSanctis andC.M. Beath (eds), Proceedings qf‘ the Twelfthlriternaiionrrl Cor$e,unce on I9$xmcrtiorl Systems.NW York. pp. 273-82.

Contractor. N.S. (I 999) ‘Self-organizing systems researchI” the social sciences: reconciling the metaphors andtlw models’, Mmmgenrent Communication Quarterly.

13: 154-66.Contractor, N. and Bishop, A.P. (2000) ‘Recontiguring

community nehvorks: the case of PrairieKNOW’, inT. lshida (ed.), Digital Ci/ies: Technologies, E.rperiertces,orld Future Pwrpectives. Berlin: Springer. pp. 15 I-64.

Contractor. N.S. and Eisenberg, E.M. (1990) ‘Com-munication networks and new media in organizations’,in J. Fulk and C.W. Steinfield (eds), Organizations andCwtunrmication Technoloe, Newbury Park, CA: Sagepp, 145-74.

Contractor, N.S. and Seibold, D.R. (1993) ‘Theoreticalframeworks for the study of structuring processes ingroup decision support systems: adaptive structurationthcoty and self-organizing systems theory’, Htrntan(‘nnwunicntiort Reseatrir, 19: 528-63.

Coniractor. N.S. and Whitbred, R. (I 997) ‘Decision develop-ment in work groups: a comparison of contingency andrclf-organizing sys tems pe r spec t ives ’ , i n G . Bamet t andL. Thayer (eds). Organizational Communication.bolerging Per:rpectirw. I/: The Rennissance in Sy.stentsr/Ii&g. Greenwich, CT: Ablex. pp. 83-104.

Contractor. N.S., Seibold, D.R. and Heller, M.A. (1996)‘Interactional influence in the structuring of media usein groups: influence of members’ perceptions of groupdecision support system use’, Human CommunicationRrrearch. 22: 45 l-8 1 .

(hlnan, M.J. and Markus, M.L. (I 987) ‘Information tech.nologies’. in F.M. Jablin and L.L. Putnam (eds)./lundhook o f Orgm~izatiorral Comrnunicatiorz: anIri/erdisciplinrrr:v Per:~perti\~e. Newbury Park. CA:Sage. pp. 42043.

Dali. R.L. and Lengcl, R.H. (1986) ‘Organizational infor-nlation requirements, media richness and structuraldesign’. Mmagernent Science, 32: 554-71.

Daly, B. (1993) ‘The influence of face-to-face versuscomputer-mediated communication channels on collec-tive induction’, Accorrrttirtg, Management & InjbnnatiorlTechnolo~, 3 (1): 1-22.

DeSanctis, G. and Poole, M. (1994) ‘Capturing the com-plexity in advanced technology use: adaptive structura-tion theory’, Organization Science, 5: 121-47.

DeSanctis, G. and Poole, M.S. (1997) ‘Transitions inteamwork in new organizational forms’, in Arhnnce.~in Group Processes, vol. 14. Greenwich, CT: JAI.pp. 157-76.

Dubrovsky, V.J., Kiesler, S. and Sethna, B.N. (1991) ‘Theequalization phenomenon: status effects in computer-mediated and face-to-face decision making groups’,Human-Computer Interaction, 6: 119-46.

El-Shinnawy. M, and Markus, M.L. (1997) ‘The povertyof media richness theory: explaining people‘s prefer-ences for electronic mail and voice mail’, IntentntionalJournal qf Human Computer Strrdies, 46: 443-67.

Farace, R.V., Monge, P.R. and Russell, H.M. (1977)Communicating and Orgnrlizirlg. Reading, MA:Addison-Wesley.

Fulk, J., Schmitz, J. and Steinfield. C.W. (1990) ‘A socialinfluence model of technology use’, in J. Fulk andC.W. Steinfield (eds), Organizations arid Com-munication Technology. Newbury Park, CA: Sage.pp. 11740.

Gallupe, R.B., Bastianutti, L.M. and Cooper, W.H. (199 1)‘Unblocking bra ins torms’ , Jownul of‘ Applied

Psychology, 76 (1): 13742.George, J., Easton, G., Nunamaker. J. and Northcraft, G.

( 1990) ‘A study of collaborative group work with andwithout computer-based support’, h~fbrmatiorr SystemsResearch, 1 (4): 394415.

Gersick, C.J.G. (1991) ‘Revolutionary change theories: amultilevel exploration of the punctuated equilibriumparadigm’, Academy of Management Review, 32:274-309.

Giddens. A. (1984) The Constitution oJSocie&c Outline @’the Throly oj’Structwation. Berkeley, CA: Universityof California Press.

Glansdorff, P. and Prigogine. I. (1971) ThermodynamicStudy o f S tructure . S tabi l i ty aitd Flrrctuatiorw.New York: Wiley.

Goodman, P.S. and Wilson, J.M. (2000) ‘Substitutesfor socialization’, in M.A. Neale. E.A. Mannix andT. Griffi th (eds), Research on Managing Groups

and Teams. Vol. 3: Technology. Stamford, CT: JAI.pp. 53-78.

Hiltz, S.R., Johnson, K. and Turoff, M. (I 986) ‘Experimentsin group decision making. I: Communications processand outcome in face-to-face vs. computerized confer-ences’, Hurnun Communicrrtion Research, 13: 225-52.

Hiltz, S.R. and Turoff, M. (1978) The Nehvork Natiou;Human Comnumicntiorz via Computw. Reading, MA:Addison-Wesley.

Hollingshead, A.B. (1996a) ‘Information suppression andstatus persistence in group decision making: the effectsof communication media’. Htrmar~ Commurlicatiorl

Research, 23: 193-219.

234 NEW MEDIA AND ORGANIZING

Hollingshead, A.B. (1996b) ‘The rank order effect:decision procedure, communication technology andgroup decisions’, Organizational Behavior and Human

Decision Processes. 68 (3): I- 13.Hollingshead. A.B. (1998a) ‘Retrieval processes in trans-

active memory systems’. Journal @‘Personality and

Social Ps)~chology, 14: 659-71.Hollingshead, A.B. (1998b) ‘Distributed knowledge and

transactive processes in groups’, in M.A. Neale,E.A. Mannix and D.H. Gruenfeld (eds), Research onh4anaging Groups and Teams, vol. 1. Greenwich, CT:JAI. pp. 103-23.

Hollingshead, A.B. (2001) ‘Communication technologies,the intemet, and group research’, in M.A. Hoggand Scott Tindale (eds), Blackwell Handbook qfSocial Psychology: Group Processes. Malden, MA.Blackwell. pp. 557-73.

Hollingshead, A.B. and McGrath, 3.E. (1995) ‘Computer-assisted groups: a critical review of the empiricalresearch’, in R.A. Guzzo and E. Salas (eds), TeamRflectiveness and Decision Making in Organizations.

San Francisco: Jossey-Bass. pp. 4678.Hollingshead, A.B., Fulk, J. and Mange, P. (2001)

‘Fostering intranet knowledge-sharing: an integrationof transactive memory and public goods approaches’, inP. Hinds and S. Kiesler (eds), Distributed Work: NewResearch on Working across Distance Using Technology.

Cambridge, MA: MIT Press.Hollingshead. A.B., McGrath, J.E. and O’Connor, K.M.

(1993) ‘Group task performance and communicationtechnology: a longitudinal shady of computer-mediatedversus face-to-face work groups’, Small GroupResearch, 24: 307-33.

Jessup, L.M. and Valacich, J.E. (eds) (1993) GroupSupport Systems: New Perspectives. New York:Macmillan,

Kiesler, S. and Sproull, L. (1992) ‘Group decision makingand technology’, Organizational Behavior and HumanDecision Processes, 52: 96-l 23.

Kiesler, S., Siegel, .I. and McGuire, T.W. (1984) ‘Socialpsychological aspects of computer-mediated communi-cation’, American Psychologist, 39: 1123-34.

Kraemer, K.L. and Pinsonneault, A. (1990) ‘Technologyand groups: assessments of the empirical research’, inJ. Galegher, R. Kraut and C. Egido (eds), Intellectual

Teamwork: Social and Technological Foundationsof Cooperative Work. Hillsdale, NJ: Erlbaum.pp. 373-404.

Lea, M. and Spears, R. (1991) ‘Computer-mediated com-munication, de-individuation, and group decision mak-ing’, International Journal of Man-Machine Studies,

34: 283-301.McGrath, J .E. (1984) Groups, Interact ion, and

Performance. Englewood Cliffs, NJ: Prentice-Hall.McGrath, J.E. (1991) ‘Time, interaction, and performance

(TIP): a theory of groups’, Small Group Research, 22:147-74.

McGrath, J.E. and Hollingshead, A.B. (1993) ‘Putting the“G” back in GSS: some theoretical issues aboutdynamic processes in groups with technological

enhancements’, in L.M. Jessup and J.E. Valacich (Group Support Sy,stems: New Perspectives. Nea IMacmillan.

rbfccGrath, LE. and Hollingshead, A.B. (1994) Gt

Interacting with Technology. Newbury Park, CA: !McGrath, J.E., Arrow, H., Gruenfeld, D.H., Hollingsl

A.B. and O’Connor, K.M. (1993) ‘Groups, tasks, Inology, and time: an integration’, Small GiResearch, 24: 406-20.

McLeod, P.L. (1992) ‘An assessment of the experimtliterature on the electronic support of group wresults of a meta-analysis’ , HumanConrk

Interaction, 7 (3): 257-80.McLeod, P.L. (1996) ‘New communication technolo

for group research: toward an integrative framewcin R. Hirokawa and MS. Poole (eds), Conrmur~icoc~~rand Group Decision Making. 2nd edn. Thousand Oa%iCA: Sage. pp. 426-62.

McLeod, P.L. and Liker, J.K. (1992) ‘Electronic mseun~systems: evidence from a low structure environmcm’.Intormation Systems Research, 3: 195-223.

McLeod, P.L., Baron, R.S., Marti, M.W. and KoII. I(1997) ‘The eyes have it: minority influence in face-to.face and computer-mediated group discussion’, J(~rrsolof;4pplied Psychology, 82: 706-l 8.

Mange, P.R. and Contractor, N.S. (2001) ‘Emergeace afcommunication networks’, in F.M. Jablin 2ndL.L. Putnam (eds), New Handbook of Organixnc~r~alC o m m u n i c a t i o n . Newbury Park , CA: Sage . pp. 410.502

Nonaka, I. and Takeuchi, H. (1995) The h’now/e~ge.

Creating Company: How Japanese Companies Crea~rthe Dynamics of‘ Innovation. New York: OxfordUniversity Press.

Poole, M.S. (1997) ‘A turn of the wheel: the case forrenewal of systems inquiry in organizational comcruni.cation research’, in G. Bamett and L. Thayer (eds),Otganizational Communication: Emerging Perspeclixf,V: The Renaissance in Systems Thinking. Greenwich,CT: Ablex. pp. 47-63.

Poole, M.S. and DeSanctis, G. (1990) ‘Understanding thruse of group decision support systems: the thee? ofadaptive structuration’, in J. Fulk and C. Steinfield(eds), Organizations and Communication Tecltrroic~,\lNewbury Park, CA: Sage. pp. 175-95.

Poole, M.S. and DeSanctis, G. (1992) ‘Microlevel sfmc-turation in computer-supported group decision msl.ing’, Human Communication Research, 19: 519.

Rice, R.E. (1984) ‘Mediated group communication’. iR.E. Rice & Associates (eds), Dze Nerv ,&II~Communication, Research, and Technology. Bevcrl)Hills, CA: Sage. pp. 129-54.

Saunders, C., Robey, D. and Vaverek, K. (1994) ‘The per.sistence of status differentials in computer conferen-i n g ’ , H u m a n C o m m u n i c a t i o n R e s e a r c h , 2 0 : 4 4 3 - 7 2

Schmitz, J. and Fulk, J. (1991) ‘Organizationalcolleaylc\.media richness, and electronic mail: a test of the somlinfluence model of technology use’, Comrtt~tnrt~crii,~!Research. 18: 487-523.

Seibold, D., Weller, M.A. and Contractor, N. tlW‘Review and critique of empirical research on ~QTU~

ORGANIZING AT THE GROUP LEVEL 235

decision support systems’, in B. Kovacic (ed.),Organizational Communiccztior~: Ne\v Perspectives.Albany. NY: State University of New York Press.pp. 143-68.

Short, J.A., Williams. E. and Christie, B. (1976) TheSocial Psychology of‘ Teleconrnrlmiccrtio,t.r. London:W i l e y .

Siegel. I., Dubrovsky, V., Kiesler, S. and MC&ire, T.W.(1986) ‘Group processes in computer-mediated com-munication’. Organizntional Behavior & HunranDecision Pmcesses. 37: 157-87.

Spears, R. and Lea, M. (1992) ‘Social influenceand the influence of the “social” in computer-mediated communication’. in M. Lea (ed.). Corlte.xtsoJ Computer-Mediated Cotnmur~ication. London:Harvester-Wheatsheaf. pp. 3&65.

Spears, R. and Lea, M. (1994) ‘Panacea or panopticon?The hidden power in computer-mediated communica-tion’, Conmmicatior~ Research, 2 1: 421-59.

Sproull. L.S. and Kiesler, S. (1991) Cowtections: NewIVqr qf Working in the Nettvorked Organization.

Cambridge, MA: MIT Press.Straw S. (1996) ‘Getting a clue: the effects of communi-

cation media and information distribution on participa-tion and performance in computer-mediated andface-to-face groups’, Snrall Group Research, 27:11542.

Straus, S. and McGrath, J.E. (1994) ‘Does the mediummatter? The interaction of task type and technology ongroup performance and member reactions’. Journal qf

Applied l?~,~ckology, 79: 87-97.Townsend. A.M., DeMarie. SM. and Hendrickson, A.R.

(1996) ‘Are you ready for virtual teams?‘, HR

Mago.ine, 41: 122-6.Trevino, L.K., Daft, R.L. and Lengel, R.H. (1990)

‘Understanding managers’ media choices: a symbolic

interact ionis t perspect ive’ , in J . Fulk andC.W. Steinfield (eds). O~gmizatiorts and Conmmicatiort

Technology. Newbury Park, CA: Sage. pp. 71-94.Valacich, J.S.. Paranka, D.. George, J.F. and Nunamaker,

J.F. (1994) ‘Communication concurrency and the newmedia: a new dimension for media richness’, Cowmunication Research, 20 (2): 249-76.

Walther, J.B. and Burgeon, J.K. (1992) ‘Relational com-munication in computer-mediated interaction’. Human

Comnrurzication Research, 19: 50-88.Wasserman, S. and Faust, K. (1994) Social Nehvork

Analysis: Methods and Applications. New York:Cambridge University Press.

Watson, R., DeSanctis, G. and Poole, M.S. (1988) ‘Usinga GDSS to facilitate group consensus: some intendedand unintended consequences’, MIS Quarterly, 12:463-78.

Weisband, S.P. (1992) ‘Group discusslon and first advo-cacy effects in computer-mediated and face-to-facedecision making groups’, Orgmimtiortal Behavior andHumm Decision Processes. 53: 352-80.

Weisband, S.P., Schneider. S.K. and Connolly. T. (1995)‘Electronic communication and social information:status salience and status differences’, Acatle,rzy q/Management Journal, 38: 112651.

Williams, E. (1977) ‘Experimental comparisons of face-to-face and mediated communication: a review’.Psychological Bulletin, 84 (5): 963-76.

Zack, M.H. and McKelmey. J.L. (1995) ‘Social contextand interaction in ongoing computer-supported man-agement groups’. Organization Science. 6: 394-422.

Zigurs, L, Poole, MS. and DeSanctis, G. (1988) ‘A studyof influence in computer-mediated group decisionmaking’, MIS Quartet+: 12: 62544.